As demand for computing power reshapes the artificial intelligence landscape, Oracle is expanding its cloud infrastructure with hardware that was once outside its core business.
The company’s decision to integrate Advanced Micro Devices’ (AMD) next generation GPUs into Oracle Cloud Infrastructure (OCI) reflects a wider change across the software sector, where control over data and performance increasingly depends on access to advanced processors.
Founded in 1977 as a database software company, Oracle built its reputation in enterprise computing and cloud services. Over the past decade it has invested in Oracle Cloud Infrastructure (OCI), which competes with Amazon Web Services, Microsoft Azure, and Google Cloud. The 2025 plan to deploy about 50,000 AMD MI450 GPUs by 2026 is an operational expansion, rather than a corporate partnership. AMD will supply the hardware, and Oracle will integrate it into its data centers to handle AI training and inference tasks.
The AI infrastructure race and economic impact
Oracle’s plan to deploy AMD GPUs drew an immediate response. AMD shares rose roughly 3% following the announcement, while Oracle stocks dipped by 1%.
The market viewed AMD’s gain as renewed confidence in its growing AI-chip portfolio and major supply deals with Oracle and OpenAI. Wolfe Research projects that data-center GPU sales could exceed USD 4 billion by late 2026. These forecasts position AMD as one of the few semiconductor firms capable of supplying cloud-scale compute hardware outside Nvidia’s ecosystem.
Oracle’s modest share decline likely reflects investor caution toward its heavy AI infrastructure spending. While the expansion strengthens its technical base and represents a long-term move to stay competitive in the AI infrastructure race, this also puts pressure on near-term margins.
Oracle: Patenting Activity
Oracle’s global patent filings remained broadly consistent between 2015 and 2022. The only noticeable dip within this period occurred in 2018, when total filings fell to 1,119, before returning to similar levels in the following years.
This consistency suggests a sustained level of innovation across Oracle’s core technology areas, supported by continuous research in data processing, networking, and enterprise software. The steady output reflects the company’s long-term approach to maintaining a balanced patent portfolio that supports its cloud and AI infrastructure strategy.
Also, it is noteworthy that around 500 patents are co-assigned with Cerner Innovation, reflecting Oracle’s integration of healthcare-related intellectual property following its acquisition of Cerner Corporation in 2022.
Oracle: Top Jurisdictions
Oracle’s patent portfolio remains concentrated in the United States, representing the core of its global protection. Beyond the U.S., the company maintains growing portfolios in China and Europe, reflecting a continued expansion into key AI and data-infrastructure markets.
Japan and India follow as active jurisdictions, likely tied to Oracle’s regional cloud operations and software-hardware integrations. Smaller but consistent activity appears in Germany, the United Kingdom, Korea, and the World Intellectual Property Organization (WO), where Oracle files international applications for later national entry.
The broad jurisdictional spread indicates a coordinated global IP strategy anchored in the U.S. but extending across Asia and Europe.
Oracle: Top Technology Areas
Oracle’s patents are concentrated in digital data processing (G06F), confirming its foundation in computing systems and cloud architecture. A significant portion also covers data transmission and networking (H04L), supporting Oracle’s cloud and database connectivity technologies.
AI-related classes are becoming more prominent, with neural networks and machine learning (G06N) and business and administrative data processing (G06Q), highlighting Oracle’s efforts to adapt traditional enterprise systems for AI workloads. Smaller but relevant clusters in wireless communication (H04W), health data processing (G16H), and computer vision and image processing (G06V and G06T) demonstrate diversification into distributed and analytics-driven applications.
Oracle: Top Law Firms
Oracle’s filings are handled by a network of U.S., Asian, and European intellectual-property firms. Kilpatrick Townsend & Stockton LLP leads, followed by CCPIT Patent and Trademark Law Office in China and Invoke. Fukami Patent Office, P.C. in Japan and Tucker Ellis LLP round out the top five.
Additional representation by Gill Jennings & Every LLP in the U.K., Hickman Becker Bingham Ledesma LLP, and Kraguljac Law Group, LLC reflects Oracle’s international filing coverage. The inclusion of The Broadgate Tower, a London office location, indicates filings managed through European patent-attorney networks. Overall, Oracle’s counsel distribution mirrors its jurisdictional reach across the U.S., Asia, and Europe.
Patents highlighting Oracle’s AI-driven security and data processing systems
Oracle’s most technically significant patents in recent years center on automating security management and applying machine learning to large-scale cloud environments. These inventions illustrate how the company’s intellectual property extends beyond enterprise databases toward adaptive, data-informed systems capable of detecting threats, analyzing behavior, and managing network activity with minimal human intervention.
AI-assisted discovery and management of application security
Managing digital security at scale requires continuous awareness of the software in use across an organization. Oracle’s engineers developed a system that automatically identifies applications within a network, evaluates their associated risks, and configures access controls to minimize threats. By assigning separate risk scores to both users and applications, the system enables dynamic restriction or approval of access based on observed behavior patterns.
U.S. Patent No. 10,536,478, titled “Techniques for discovering and managing security of applications,” was filed on February 23, 2017, and granted on January 14, 2020. The patent was represented by Kilpatrick Townsend & Stockton LLP, with inventors Kirti Ganesh, Kamalendu Biswas, Sumedha Nalin Perera, and Adina Florina Simu. Furthermore, this patent is Oracle’s most cited patent, reflecting its influence within the company’s intellectual-property portfolio, and is classified under digital communication.
Cloud-based security monitoring through pattern recognition and deep learning
With the growth of distributed cloud services, detecting abnormal user behavior has become critical to cybersecurity. This invention describes a system that collects user-activity data from cloud environments, learns typical behavioral patterns through unsupervised models, and identifies deviations that may signal security breaches. The approach allows large-scale monitoring without predefined rules, reducing false positives and adapting to new threat patterns over time.
U.S. Patent No. 11,165,800, titled “Cloud based security monitoring using unsupervised pattern recognition and deep learning,” was filed on May 25, 2018, and granted on November 2, 2021. The patent was represented by Kilpatrick Townsend & Stockton LLP and lists Sajjit Thampy as inventor.
Machine-learning processing of sequential interaction events
Understanding user interaction sequences is essential for accurate profiling and anomaly detection in online systems. This patent presents a technique for analyzing ordered event data tagged within web or app environments. Using a machine-learning model, the system interprets these tagged sequences to infer user profiles or behavioral states, enabling more precise responses to user activity in real time.
U.S. Patent No. 10,938,927, titled “Machine learning techniques for processing tag-based representations of sequential interaction events,” was filed on August 30, 2017, and granted on March 2, 2021. The patent was represented by Kilpatrick Townsend & Stockton LLP and lists Peter Crossley and Ethan Dereszynski as inventors.
